Journal of Environmental Planning and Management,
Vol. 47, No. 5, 737–755, September 2004
Nitrogen Input from Residential Lawn Care Practices in
Suburban Watersheds in Baltimore County, MD
NEELY L. LAW*, LAWRENCE E. BAND* & J. MORGAN GROVE†
*Department of Geography, University of North Carolina, Chapel Hill, NC, US
†Department of Agriculture Forest Service, South Burlington, VT, USA
(Received July 2003; revised February 2004)
ABSTRACT A residential lawn care survey was conducted as part of the Baltimore
Ecosystem Study, a Long-term Ecological Research project funded by the National
Science Foundation and collaborating agencies, to estimate the nitrogen input to urban
watersheds from lawn care practices. The variability in the fertilizer N application rates
and the factors affecting the application rates were examined. Results indicated that the
annual input of nitrogen from fertilizer is a major component of the urban watershed
nitrogen budget and it is both spatially and temporally variable. There is a wide range
in the application rate of fertilizer N to residential lawns applied by homeowners and by
professional lawn care companies. Survey data estimated a mean fertilizer application
rate of 97.6 kg N/ha/yr with a standard deviation of 88.3 kg N/ha/yr. Analyses suggested
that the fertilizer application rate is affected by social economic factors and soil
characteristics to include the market value of the house, age of development, soil bulk
density and soil nitrogen content.
Introduction
The lawn has been a fascination and pre-occupation with American homeowners
for over a century and the lawn care industry in the USA has grown to be a
multi-billion dollar industry (Teyssot, 1999). In the State of Maryland, lawn care
companies, followed by residents of detached single-family homes, provided the
greatest economic support to the turfgrass industry (Maryland Turfgrass Council, 1996). Research has shown that nitrogen application rates of fertilizer to
residential lawns are comparable to fertilizer applications in agriculture and golf
courses (e.g. Gold et al., 1990; Liu et al., 1997). However, the impact of fertilizer
applications to lawns on the nitrogen cycle in urban watersheds is not well
understood. Given current mandates to reduce nutrient loadings to surface
Correspondence Address: Neely L. Law, Department of Geography, University of
North Carolina, Chapel Hill, NC 27599, USA. Tel: 919 962 8901, Fax: 919 962 1537.
E-mail: nlaw@email.unc.edu
0142-6397 Print/1469-9710 Online/04/050737-19 © 2004 University of Newcastle upon Tyne
DOI: 10.1080/0964056042000274452
738 N. L. Law et al.
waters such as the Chesapeake Bay, a better understanding of the sources and
pathways of nitrogen is needed to achieve this water quality goal.
Run-off from residential lawns and contaminated groundwater are two
types of non-point source pollution in urban watersheds that affect surface water
quality. It is well documented that groundwater from agricultural land-use
practices has elevated nitrate concentrations from the long-term use of fertilizers.
The conversion of agriculture to urban land uses typically results in a reduction
in nitrate concentrations and loads to surface water (Band et al., 2002; Weller et
al., 2003). However, the legacies of past land-use practices, coupled with current
residential fertilizer input, create a challenge to determine how best to manage
non-point source pollution from urban watersheds.
Fertilizers contain two of the major nutrients, N and P that contribute to
water quality impairment in the surface waters of the USA, including the
Chesapeake Bay (Vitousek et al., 1997; US Geological Survey, 1999). Plot-scale
studies have shown that nitrate in run-off from fertilized turfgrass is not
necessarily a constant source of nutrients to surface water but is temporally
variable. For example, field studies have found low concentrations of nitrate in
storm run-off from fertilized lawns (Petrovic, 1990; Garn, 2002), with the
exception of storm events after fertilizer applications (Kelling & Peterson, 1975).
For example, Line et al. (2002), working in suburban sites in North Carolina,
measured nitrogen concentrations to be four times higher in run-off from single
family residential land uses during February compared to fall or winter concentrations. Seasonal variability has also been observed for soil water nitrate
concentrations from fertilized lawns. For example, Gold et al. (1990) measured
soil-water nitrate concentrations in fertilized residential lawns in Rhode Island to
be approximately seven times greater in the spring (e.g. 2.6 mg/L) compared to
other seasons. Other studies in Rhode Island and Michigan found higher
concentrations of nitrate-N in soil water in the fall compared to spring (Miltner
et al., 1996; Liu et al., 1997).
As turfgrass is a dominant land cover type in urban watersheds with a
majority of homeowners fertilizing their lawns (Center for Watershed Protection,
2000), there is concern regarding the impact of lawn fertilizers on water quality
(Waschbusch et al., 1999; Robbins et al., 2001; Garn, 2002). In 1996, a survey by
the Maryland Turfgrass Council estimated that 19% of the state area is covered
by grass, inclusive of sod farms, golf courses, institutional land uses, parks etc.
and 56% of this area is associated with single family residential land use. With
the continued expansion of urban areas and residential development, fertilizer
applications will represent an increasingly important source of nitrogen to the
watershed (Maryland Department of Planning, 2001; Robbins et al., 2001).
This paper describes the results of a residential lawn care survey which was
conducted as part of the Baltimore Ecosystem Study (BES), an urban Long Term
Ecological Research (LTER) project funded by the National Science Foundation and collaborating agencies. The survey was carried out to estimate the
amount of nitrogen applied to urban watersheds from residential lawn care
practices. The BES is one of two urban LTER sites that studies urban areas as
ecological systems, where the components of the ecosystem extend to include
the behaviours, relationships and interactions of people at varying scales (i.e.
individuals, households and neighbourhoods) with other components of the
ecosystem. The survey will contribute data to address one of the major BES
research questions: ‘What are the fluxes of energy and matter in urban ecosys-
Nitrogen Input in Watersheds from Residential Lawn Care
739
tems and how do they change over the long term?’ (http://www.beslter.org).
An initial study of nitrogen soil dynamics from various land cover types found
that nitrogen cycling in lawns may be high, but losses may be low due to high
soil organic matter content and active retention (Groffman et al., 2002). Data
generated from the survey, in addition to future modelling work, will determine
the significance of specific sources of nitrogen that may contribute to non-point
source pollution to surface water in urban areas. The specific objectives of the
survey were to:
• Determine the temporal, spatial and socio-economic distribution of fertilizer N
in urban watersheds;
• Identify the factors, both socio-economic and ecological that affect fertilizer N
application rates;
• Estimate the contribution of lawn fertilization to the watershed nitrogen
budget; and
• Define lawn fertilization rates as parameters for use in hydro-ecological
models.
Study Site Description
The survey focused on two watersheds, Glyndon and Baisman Run that are a
part of the BES monitoring network within the Gwynns Falls and Gunpowder
watersheds, respectively (Figures 1 and 2). The Glyndon catchment has an area
of 0.8 km2 and includes a portion of an historical colonial town (Reisterstown).
It is a headwater catchment of the Gwynns Falls watershed. Glyndon is characterized by gentle slopes producing diffusive flow patterns in the upper part of
the watershed. Soils in the catchment are largely classified as loam and silt loam
that are moderate to strongly acidic (USDA, 1976). The effect of development
within the catchment is physically apparent from the engineered drainage
networks from past agricultural practices and the incised main channel from
more recent urban development. The predominant land use is residential, served
by public water supply and sanitary sewers, with a mix of other urban and open
space uses (e.g. recreational field) (Table 1). Downstream of Glyndon, the
Gwynns Falls River drains more intensive urban land use moving from Baltimore County into Baltimore City.
Baisman Run is part of the Gunpowder water supply catchment and is
characterized by low density, large lot development on septic systems in the
upper third of the watershed. Downstream, the remaining two-thirds of the
watershed is completely forested and is part of the Baltimore County parks
system. Similar to Glyndon, the morphology of the stream network has been
altered from human-influenced land-use practices. Hillslope gradients proximal
to the main branch of Baisman Run are typically steeper than hillslopes in the
lower order tributaries. Slopes are convex and range from 0–5% at the ridgetops
and up to 40% at the lower hillslopes. Soils within the catchment are classified
as loam and silt loam and are generally acidic (USDA, 1976).
Survey Design and Methods
A household survey was designed to collect data on lawn management practices
in urban-suburban watersheds to determine how much fertilizer the average
homeowner uses per application and annually and how fertilizer and outdoor
740 N. L. Law et al.
Figure 1. The Baltimore Ecosystem Study watersheds and stream network for the Gwynns Falls and
its subwatersheds, Baisman Run and Pond Branch watersheds. The Glyndon watershed is located in
the headwaters of the Gwynns Falls watershed.
water use practices vary within urban watersheds. Survey data were generated
at this fine spatial scale (i.e. parcel) to associate lawn care practices with
socio-economic and other parcel data from the Maryland Property View© dataset
to include: the tax assessed market value of the house (land value plus improvement value), lot size, and square footage of the house.
The survey was administered door-to-door and comprised of two parts: (1)
a set of 23 questions and (2) with the permission of the homeowner, soil samples
from the lawn were taken to estimate soil bulk density and provide information
on general soil chemistry. The survey was designed to determine fertilizer
application rates and water use practices. A limited set of additional household
and property characteristics were recorded to include type of grass, proportion
of weeds and other landscaping features (e.g. permeable area other than
turfgrass) and built structures, including driveways and sidewalks. A distance
Nitrogen Input in Watersheds from Residential Lawn Care
741
Figure 2. Land cover within the study watersheds of Glyndon and Baisman Run from high
resolution digital colour infrared aerial photographs (October 1999). The outline of the watershed
boundaries is indicative of vegetation land cover types to include lawns.
measuring wheel was used to estimate the lawn area for property in the
Glyndon watershed. Lawn area in Baisman Run was estimated based on visual
surveys that approximated the percent area of the property that was turfgrass,
landscaped and had built structures, including driveways and sidewalks. These
estimates were corroborated with information from the Maryland Property
View© dataset on the lot size and building footprint, as well as high resolution
742 N. L. Law et al.
Table 1. Summary characteristics of land cover land use, population
density and housing development in the study watersheds
Watershed area (km2)
Residential
Forest
Open urban space
Commercial, institutional
Percent lawn area
Population density (pers/ha)
Housing density (house/ha)
Average square footage of house
Average lot size (ha)
Glyndon
Baisman Run
0.8
47%
4%
16%
32%
15%
9.4
3.9
1534
0.13
3.7
34%
66%
0
0
25.5% (75%)1
1 (3.0)
0.3 (1.0)
3082
0.93
Note: 1Values in parentheses refer to the residential portion of the Baisman Run watershed
and not the whole watershed.
digital orthophotography. A follow-up survey to the professional lawn care
companies identified in the household survey was conducted to document
industry lawn care management practices.
The study watersheds were partitioned into subdivisions and households
selected to participate in the survey using a GIS framework. Subdivision
boundaries were defined using high resolution (1m) digital colour infrared aerial
photographs, an overlay of the road network, parcel data attributed with the
year the house was built from Maryland Property View© and watershed
boundaries. Six subdivisions were digitized within the Glyndon watershed, and
four subdivisions within Baisman Run. Only residential, homeowner-occupied
households were identified to participate in the survey. Nine of the subdivisions
were single family detached homes and one subdivision was a townhouse
development. A random starting house was selected using GIS software ARCView3.2a. A total of 10 households were surveyed per subdivision for a sample
size of 60 in Glyndon and 40 in Baisman Run.
Composite soil chemistry was based on three to five randomly selected soil
cores of equivalent depths ranging from 3 to 6 cm. The soil was dried and
processed through a 0.5mm sieve prior to analysis at the University of Maryland
Soil Testing Laboratory. Soil carbon, hydrogen and nitrogen (to include organic)
were measured using a Leco CHN 2000 Carbon-Hydrogen-Nitrogen Analyzer.
The instrument has a detection range of 0.01% to 100% for carbon and nitrogen.
A soil corer was used to take samples from the top 5 cm of the soil at three
randomly selected sites from the lawn to measure average bulk density. Given
the purpose of the study, analyses focused on soil nitrogen as an input to
residential watersheds from lawn fertilizer practices. Other BES studies provide
more complete analyses of soil quality in urban watersheds (e.g. Pouyat et al.,
2002; Pouyat et al., 2004).
An average application rate of N (kg N/ha/yr) was calculated at three
spatial scales from the data generated from the household and professional lawn
care company surveys, in addition to land use-land cover characteristics within
each watershed. The application rate estimated from survey responses on the
type and amount of fertilizer product, the frequency of application and lawn
area. A commonly used fertilizer formulation (29N-3P-4K) was used to estimate
Nitrogen Input in Watersheds from Residential Lawn Care
743
Table 2. Summary of survey results aggregated to the catchment
level.Values in parentheses indicate ⫹ / ⫺ one standard deviation from the mean watershed annual application rate
% Fertilize
% Self-apply
% Profession apply
Kg N/ha/yr
Soil tested (%)
Water lawn
Glyndon
Baisman Run
68
71
29
12.5 (10.2–14.4)
56
44
56
9.5 (7.7–11.2)
27.8 (22.6–33.0)
13
9
19
48
the fertilizer N application rate for 56% of the surveys. This estimation method
was reasonable as respondents provided information on the frequency and
amount of fertilizer used, in addition to the product name, but the specific
product type (e.g. N-P-K formulation) was not identified by the survey respondent. The percentage of homeowners who use a lawn care service or fertilize
their own lawn within each watershed was used to provide a weighted average
application rate. The weighted average application rate multiplied by the lawn
area within each watershed provided a total mass or input of nitrogen applied
to lawns. The mass of nitrogen, divided by the watershed area (mass per unit
watershed area, mpw), residential land use area (mass per unit residential area,
mpr) and lawn area (mass per unit lawn area, mpl) provides an application rate
of fertilizer at three different spatial representations of watershed land use-land
cover characteristics. The mass per unit watershed area, mpw, is a useful
measure to develop nutrient mass balances at the watershed scale for ecological
and hydrological studies.
The survey information, soil characteristics, and data on the socio-economic
characteristics of the subdivisions were used in linear regression analyses to
determine the controlling factors that may affect nitrogen input to the watersheds from lawn care practices. The average application rate (Kg N/ha/yr as
mpl) for each subdivision was used as the unit of analysis and is based on the
individual homeowner application rates. Individual household responses were
not used given the unequal number of survey questionnaire responses and soil
data for each household. Further, the methods used to estimate the application
rates are well constrained by the lawn area, but there may be significant
household level errors given the self-reporting method used in the survey.
Results and Discussion
Survey Results of Lawn Management Practices and Soil Analyses
Overall, the survey response rate was high with 68% and 80% of the residents
in Glyndon and Baisman Run respectively, for a total sample size of 73
households. A lower percentage of homeowners permitted soil samples to be
taken with 28 soil samples taken in Glyndon and 27 in Baisman Run. A
summary of survey data is presented in Table 2. The survey found that a
majority of homeowners in both watersheds fertilized their lawns with 68% of
744 N. L. Law et al.
the residents in Glyndon applying fertilizer to their lawns and 56% of homeowners in Baisman Run. This is consistent with survey findings by the Center for
Watershed Protection (2000) that found approximately 70% of all lawns are
regularly fertilized. In Baisman Run, more residents employed a professional
lawn care service compared to residents in the Glyndon watershed. The frequency of application by homeowners is typically one or two times a year in
spring and/or fall and up to six applications a year for residents that employ a
professional lawn care service. However, the greater frequency of applications
did not necessarily translate into a higher annual application rate of nitrogen
given the different fertilizer products used on a seasonal basis by the professional lawn care companies. Homeowners reported using the same product
throughout the year, whereas professional lawn care companies reported using
less nitrogen per application during the spring and summer compared to the fall
applications.
The average application rate of fertilizer on a lawn area basis reported by
professional lawn care companies is higher compared to that calculated for
homeowners. However, both are within the range recommended by the Maryland Cooperative Extension Service guidelines. At the scale of individual lawns,
the survey of the professional lawn care companies found that the percentage of
N in the product varied seasonally and that the average annual application rate
ranged from 100.1 to 161.1 kg N/ha/yr of lawn (2.1 to 3.3 lbs N/1000 ft2 of
lawn). This application rate is based on four to six applications per year.
However, some homeowners surveyed only use a professional lawn care service
one to three times year, thus reducing the actual annual application rate by
professional lawn care companies. The average annual application rate by
homeowners who apply their own fertilizer was 106.9 kg N/ha (2.2 lbs N/1000
ft2) with a standard error of 21.8 kg N/ha and a range from 10.5 to 369.7 kg/ha
(0.2 to 7.6 lb N/1000 ft2). The higher application rates were associated with the
newest development of single family homes and the townhouse development. A
large range in application rates was also found in a North Carolina study
(Osmond & Platt, 2000). A list of fertilizer application rates reported in the
literature is provided in Table 3. The values represented in the Table were used
largely for experimental studies and represent an end range in application rates
and are not necessarily characteristic of residential lawn care practices.
Differentiation in the application rate of fertilizer between the two watersheds is summarized at three spatial aggregations, reflecting differences in land
cover, development type and household behaviour between the two watersheds
(Table 4). A higher application rate was estimated within the Glyndon watershed, compared to Baisman Run. At the watershed scale, the difference in the
application rate is largely a function of the area of the watershed established by
the location of the stream gauge. Hence, the impact of the development in
Baisman Run as measured by the gauge is reduced due to mixing with the
cleaner water from the forest cover of the lower two-thirds of the catchment and
in-stream processes (Law, 2003). If the residential land area is only considered,
the application rate for Baisman Run and Glyndon are similar, indicative of an
averaging effect between the higher application rate and smaller lawn area in
Glyndon and the lower application rate and larger lawn area in Baisman Run.
The greatest distinction between application rates is estimated when only the
lawn area is considered within each of the watersheds. The application rate in
Glyndon increased to 83.5 kg N/ha/yr, mpl compared to 37.1 kg N/ha/yr, mpl
Nitrogen Input in Watersheds from Residential Lawn Care
745
Table 3. Summary of fertilizer application rates in other studies
Reference
Application rate
Comments
Study location
Kelling & Peterson
(1975)
49 kg N/ha
298 kg N/ha
225 kg N/ha
Homeowner applied
Fall application
Considered high/
excessive
9 urban lawns (locations
not given)
Flipse et al. (1984)
Starr & DeRoo (1980)
107 kg N/ha/yr
180–195 kg N/ha/yr
Liu et al. (1997)
149 kg N/ha/yr
Garn (2002)
146–171 kg N/ha/yr
King et al. (2001)
49–540.9 kg N/ha
Morton et al. (1988)
97 kg N/ha/yr
Miltner et al. (1996)
244 kg N/ha/yr
196 kg N/ha/yr
Erickson et al. (2001)
50 kg N/ha
NCSU Water Quality
Group (2000)
Petrovic (1990)
29–151kg N/ha
(average)
0–2148 kg N/ha
24–224 kg N/ha
Long Island, NY
2 applications per year Windsor, Connecticut,
field plots
Moderate application Kingston, Rhode Island,
rate divided
field plots
equally amongst 3
applications per year
Assumed residents
Lakeshore lawns in
followed manufacturers Walworth County,
recommended rates,
Wisconsin
4 applications per year
Range of average
Golf course in Austin,
reported application
TX
rates for roughs,
fairways, tees
and greens
Low
Kingston, Rhode Island,
field plots
High
5 applications of
Michigan State
39.2 kg N/ha
University, field plots
Moderate application University of Florida,
rate
field plots
Based on household
4 North Carolina
survey data
communities
CWP (2000)
49 kg N/ha/yr
267 kg N/ha/yr
97.5–195 kg N/ha/yr
Based on compilation
of application rates in
a literature review
Prior to 1940
1970s
Current extension and
garden literature
recommended rates
in Baisman Run. At the spatial aggregation of the lawn, the application rate is
more likely to be associated with individual homeowner behaviour as residents
of Glyndon have, on average, a higher fertilizer application rate compared to
residents of Baisman Run.
Significance of Lawn Fertilization to Catchment N budget
Based on the data at the watershed level, Groffman et al. (2003) estimated that
approximately 53% of the total nitrogen input for the Glyndon watershed is
from the application of fertilizers with the remainder from atmospheric deposition. In Baisman Run, additional inputs of nitrogen from estimates of septic
746 N. L. Law et al.
Table 4. Application rate of fertilizers to lawns at three spatial
scales
Application rate (kg N/ha/yr)
Scale1
Watershed (mpw)
Residential land use (mpr)
Lawn cover (mpl)
Glyndon
12.5
26.7
83.5
Baisman Run
9.5
27.8
37.1
Note: 1The spatial scale is an expression of the aerial aggregation of the watershed area, total residential land-use area within each watershed and total lawn
area within each watershed. Each different scale is abbreviated as mass per unit
watershed, mass per unit residential land use or mass per unit lawn area as,
mpw, mpr, and mpl, respectively.
system effluent load decrease the contribution of nitrogen from fertilizer to
approximately 35%. Compared to the measured output in surface water, there is
considerable retention and/or loss within the watersheds. The watershed budgets do not consider leakages from sanitary sewer infrastructure in Glyndon and
use conservative estimates of the total N septic tank effluent based on data
presented by Swann (2001). There were no intentional discharges within each of
these watersheds from commercial or industrial practices. Groffman et al. (2003)
suggest that these budgets are an underestimation of N input and discuss the
associated uncertainty of these measurements in greater detail.
Household Irrigation Patterns
Another lawn care practice surveyed was lawn watering. Overall, 32% of the
survey respondents watered their lawn. The majority of lawn watering occurred
in the Glyndon watershed. Only 13% of the survey participants in Baisman Run
watered their lawn. This reflects the fact that Baisman Run residents use private
wells for domestic water supply and at the time of the survey, summer water
deficits and drought conditions existed. As a consequence, water use for exterior
purposes was limited in Baisman Run.
Soil Characteristics and Testing Behaviour
The last part of the survey included soil samples to be taken from the lawn for
soil bulk density and chemistry. It was found that the average soil pH of the
residential lawn is within an acceptable range for healthy turfgrass as recommended by the Maryland Cooperative Extension Service, although soils in the
area are generally slightly acidic (USDA, 1976). The carbon content of the soils
is within the low range for the Oxisols soil order, but is generally deficient in soil
N (Brady, 1990). A summary of general soil chemistry for the two watersheds is
provided in Table 5. Of the survey respondents, 17% had their soil tested to
determine the most appropriate N-P-K ratio in a fertilizer to apply to their lawn.
This corresponds with data cited by the literature review conducted by the
Center for Watershed Protection (2000), where it was found that less than 20%
of residents had their soil tested prior to the application of lawn fertilizer. In the
Soil type
Silt loam
Silt loam
Watershed
Glyndon
Baisman Run
1.3 (.21)
1.2 (0.15)
Bulk density (g/cm3)
6.8 (0.6)
5.9 (0.54)
pH
3.3 (1.0)
4.4 (0.93)
% Organic matter
1.6 (0.98)
1.7 ( 0.51)
% Carbon
0.03 (0.01)
0.04 (0.01)
% Nitrogen
0.7 (0.09)
0.8 (0.11)
% Hydrogen
Table 5. Summary of soil properties and chemistry from residential lawns within the study watersheds. The average values for the
soil samples are provided with the standard deviation given in parentheses
Nitrogen Input in Watersheds from Residential Lawn Care
747
748 N. L. Law et al.
Figure 3. The mean application rate of fertilizer per unit lawn area as a function of median age of
house within each subdivision.
household survey, it was found that a greater percentage of residents who use
a professional lawn care company had their soil tested. The practice of soil
testing is likely to increase with the passage of legislation in Maryland (Code of
Maryland Regulations 15.20.06) that requires professional lawn care companies
to test the soil for nutrient management purposes every three years.
The Influence of Socio-economic Factors
Statistical analyses found a significant, positive linear relationship between the
average application rate of nitrogen by subdivision (kg N/ha/yr, mpl) and the
median year the house was built within each subdivision (r2 ⫽ 0.45, pvalue ⫽ 0.04) (Figure 3). The relationship suggested a set of hypotheses regarding a greater application rate for the more recently developed homes, which is
a function of:
• higher socio-economic status of the newer homes; the newer homes have a
higher market value; or
• more fertilizer is being applied to help establish lawns due to the poor soil
quality from recent construction.
It would be expected that a higher application rate of fertilizer, (i.e. consumption
of resources) would be associated with higher socio-economic status. A strong
correlation existed between the market value of the house and lot size (e.g.
r2 ⫽ 0.75), but lot size did not significantly explain the variation in the average
N fertilizer application rates between subdivisions or the likelihood of homeowners to fertilize their lawn. A mixed-effects relationship emerged between the
nitrogen application rate of fertilizer and the market value of the house (Figure
Nitrogen Input in Watersheds from Residential Lawn Care
749
Figure 4. The mean application rate of fertilizer as a function of the average market value of houses
within each subdivision. In this plot, point ‘A’ is a townhouse development while point ‘B’
represents the newest subdivision.
4). The data suggest that households with intermediate housing values, relative
to the survey sample, have the highest application rate of fertilizer N and that
there is similar behaviour in lawn management practices, as per fertilizer
application, for the end ranges of social-economic households. This is consistent
with a trend that emerged from Osmond & Platt (2000), where middle-income
level households had the highest application rate of N fertilizer. The points
positioned outside the trend in Figure 4 may be explained by the distinct
development patterns within these subdivisions. The ‘A’ subdivision is a townhouse development, unlike the single family detached development in all other
subdivisions surveyed and point ‘B’ is the more recently constructed subdivision. The point ‘C’ has a disproportionately high percentage of professional
lawn care companies that fertilize lawns within this subdivision, relative to other
subdivisions in the survey, but may also reflect an upswing in the form of the
relationship as suggested in Vogel (1999) (see discussion below). Within this
curve, points below and above $200 000 are representative of the two different
catchments (Figure 4). The points on the ascending limb of the curve are the
Glyndon subdivisions and the points on the descending limb of the curve are the
subdivisions in Baisman Run. Therefore, a misleading interpretation may be
taken if data from only one watershed was surveyed.
The form of the relationship in Figure 4 reflects, in part, the properties of a
Kuznet curve which is an inverted U-shaped curve that describes the effect of
per capita income on environmental quality (Vogel, 1999). This theory states that
pollution increases as income rises but reaches a threshold value where pollution
levels begin to decline as a threshold income value is exceeded. Although the
form of the relationship for the survey data is similar to the environmental
750 N. L. Law et al.
Kuznet curve, the Kuznet curve is typically based on larger spatial scales than
households or subdivisions (e.g. countries). The application of such an analysis
to smaller, more local scale studies may warrant further research to better
understand how social and ecological drivers within the suburban landscape
impact ecosystem dynamics at this fine scale. The form of this relationship, taken
with the strongly linear relationship of fertilizer application rates and the
median house age within each subdivision (Figure 3), suggests that household
behaviour for lawn management can be influenced and changed with different
socio-economic settings. Lawn management behaviour within American suburban neighbourhoods is strongly influenced by community norms and pressure,
reinforced by the desire to gain acceptance within the neighbourhood and desire
to reflect a particular status-level (Teyssot, 1999; Robbins et al., 2001). The
positive social feedback created by the community behaviour will have implications on the watershed ecology in terms of the inputs and outputs of nutrients
and energy on turfgrass (Grove et al., 2003). These analyses, therefore do not
support Hypothesis I, that fertilizer application increases with higher socio-economic status.
The Influence of Physical and Ecological Factors
The physical and chemical properties of the soil influence the quality of the
turfgrass as the soil environment will partly control the moisture and nutrient
availability for the grass. A difference of means test (t-test) indicated a significant
difference in soil bulk densities between the two watersheds (p-value ⫽ ⬍ 0.01)
with Glyndon having higher bulk densities compared to Baisman Run (Table 5).
More recently developed subdivisions have higher bulk densities compared to
the older, more established lawns (Figure 5). Additional correlation analyses as
summarized in Table 3, further support Hypothesis II that the more recently
developed subdivisions have poorer soil quality. For example, there is a negative
relationship between the ‘year built’ variable with ‘percent soil carbon’ and
‘percent soil nitrogen’ that indicates more recently developed homes have lower
soil N and C content compared to older development sampled.
The higher bulk densities in the more recently developed properties are
attributed to the more compacted nature of the soil from recent construction
traffic and fill material as suggested by Hamilton & Waddington (1999). Some of
the compaction may also be attributed to lawn care practices by pedestrian and
machine traffic (Kelling & Peterson, 1975). The survey data found a significant,
positive relationship between bulk density and the average annual fertilizer
application rate by homeowners within the subdivisions (r2 ⫽ 0.75, pvalue ⫽ 0.001).
The compacted nature of the soil in the more recently developed properties
may reduce the leachate potential of nitrate-N into the soil profile and shallow
groundwater, but may increase the run-off potential from the lawn. Higher bulk
densities associated with the more recently developed lots is consistent with
Legg et al. (1996), who used age of lawn, or years since development of lawn, as
a surrogate for many soil properties in a study of run-off from residential lawns.
Soil structure and macropores develop as a lawn age and the accumulation of
organic matter and bioturbation as fauna develop decreases the bulk density of
the soil and increases infiltration rates. Further, the more recently developed
homes with higher bulk density from recent construction activity and fill
Nitrogen Input in Watersheds from Residential Lawn Care
751
Figure 5. Average soil bulk density (g/cm3) by the median age of house within each subdivision for
the Baisman Run and Glyndon watersheds.
material apply greater amounts of fertilizer. Based on experiments by Legg et al.
(1996), it would be expected that the run-off potential would be greater for the
newly developed lawns with higher bulk density. The covariation in soil
hydraulic properties such as lower infiltration rates for higher bulk density soils
with higher fertilizer application rates could generate higher concentrations of
nitrate and phosphate in storm run-off compared to the more established lawns.
Fertilizer application rates are correlated with soil N (p-value ⫽ 0.08,
r2 ⫽ 0.33) (Figure 6) and it was found that soil with higher N content had lower
application rates than soil with lower N content. The soils with the lower soil N
and soil C content are also associated with the more recently developed
subdivisions (see Table 6). The relationship between the application rate of
fertilizer and soil N content can be explained by the ‘smart farmer hypothesis’
as higher soil N content would promote healthier turfgrass. Given that few
homeowners had their soil tested, it is suggested that homeowners are relying
more upon qualitative or visual assessments of their lawn to determine the
amount, or if any fertilizer is applied. Turfgrass with a sufficient nitrogen pool
for shoot growth will appear healthier than lawns with lower soil N content. The
higher soil carbon content in the lawns surveyed, with the greater soil N content
(r2 ⫽ 0.86) is supportive of the fact that soil quality may affect the amount of
fertilizer that is applied to residential lawns. Further, Petrovic (1990) illustrated
that total N in the surface soil layer (0–10cm) increases with the age of the
turfgrass and asymptotes at 20 years and that the greatest change in soil N
752 N. L. Law et al.
Figure 6. Fertilizer application rate as a function soil N content.
Table 6. Correlation coefficients between subdivision
age and tax value, with soil characteristics averaged for
the Glyndon and Baisman Run watersheds
Soil bulk density
Percent organic matter
pH
Percent soil carbon
Percent soil nitrogen
Year built
Market value
0.83
⫺ 0.23
0.28
⫺ 0.59
⫺ 0.41
⫺ 0.1
0.62
⫺ 0.64
0.02
0.45
occurs within the first 10 years of turfgrass establishment. The age of development within the subdivisions with the lower soil N content and higher application rates of fertilizer were less than 10 years old at the time of this survey. The
health of the lawn is also contingent upon other factors, such as soil moisture or
disease stress. Thus, an incorrect assessment of the factors leading to the
unhealthy appearance of turfgrass may lead to inappropriate decisions on lawn
care management practices. Therefore, the analyses do support Hyopthesis II,
that more fertilizer is being applied to lawns due to poor soil quality from recent
construction.
Conclusions
The residential lawn care survey is one of a few studies that attempted to
determine the input of nitrogen from fertilizers within an urban watershed. It
was found that approximately 53% of the total nitrogen budget in Glyndon is
Nitrogen Input in Watersheds from Residential Lawn Care
753
from lawn fertilization. Although the input of fertilizer N accounted for a large
component of N input to the watershed, survey results found that the application rates of fertilizer to individual lawns are comparable to findings in other
studies and recommendations by the Maryland Cooperative Extension Service
on lawn care (e.g. 97.7–195.3 kg N/ha of lawn or 2–4 lbs of N per 1000 ft2),
depending upon grass species. However, the spatial scale at which the application rates are estimated can yield large variation in the results. Given varying
spatial representations of land use-land cover within the watershed, a distinction
between the two watersheds emerged.
Within a single land use type, residential, the watershed-based survey
indicated that there is a spatial and temporal pattern to nitrogen input from
residential lawn care practices that is influenced by socio-economic and soil
characteristics. The application rate of fertilizer varied spatially, by subdivision
where greater rates of application were associated with the age and type of
development. The temporal distribution of fertilizer input, however, was similar
across subdivisions that generally occurred in the spring and fall but varied in
the frequency of applications. The form of the relationships and statistical
significance suggested that a subdivision’s socio-economic status will influence
the input of N fertilizer to residential lawns, in a non-monotonic function, rather
than linear, as initially hypothesized. A higher application rate of fertilizer is not
linearly associated with higher socio-economic status. A more linear relationship
was found to exist, however, between variables indicative of soil quality and
fertilizer N application rates, where more fertilizer was being applied to lawns
with poorer soil quality.
The potential consequence of the variable management practices within the
same residential land use type is the generation of hot spots for nutrient run-off.
Based on the statistical analyses, it is suggested that sites with the greatest
potential for nitrate export in run-off are properties trying to establish a healthy
green lawn within suburban watersheds (i.e. recent development). To minimize
this impact, soil testing of lawns may need to go beyond soil chemistry and
include physical characteristics such as bulk density or other measures or indices
that indicate the run-off potential from lawns.
Future applications and extensions of the household survey will be used to
explore how behavioural patterns affect nutrient input to watersheds from
fertilizer use. This information would be used to further define the potential
feedback system that was suggested by the initial study where a change in
fertilizer practices may result given a change in socio-economic setting. For
example, data are needed on a wider distribution of development types and
socio-economic characteristics. More detailed information on social demographic
characteristics (e.g. income) and attitudes/perceptions of lawn care management
would provide greater insight toward behavioural patterns associated with
fertilizer N input within urban watersheds. Additional survey questions would
address the attitude of homeowners towards the importance of the characteristics of neighbours’ lawns, the perceived impact the quality of lawn has on
property value, and environmental impact of fertilizer on water quality. Future
work will also include the use of a hydro-ecological simulation model to
determine the impact that lawn fertilization has on nitrogen cycling and output
within urban, residential watersheds. Model simulations will be used to explore
how a change in the amount of fertilizer applied, when and where it is applied
within the watershed affects surface water quality. A coupled understanding of
754 N. L. Law et al.
the sources of nitrogen and the hydrologic flowpaths within residential watersheds would provide an improved understanding of non-point source pollution
dynamics and lead to more focused efforts to reduce its impact on water quality.
Acknowledgements
This research is supported by the Baltimore Ecosystem Study project, National
Science Foundation Long-Term Ecological Research program, grant number DEB
9714835, and by the EPA-NSF joint program in Water and Watersheds, project
number GAD R825792. We thank the USDA Forest Service Northeastern Research Station for site management, and in kind services to the BES. In addition
we thank the University of Maryland, Baltimore County for their contribution to
office and laboratory space at the Research Technology Center on their campus.
The City of Baltimore Department of Parks and Recreation and Department of
Public Works, the Baltimore County Department of Parks, the Maryland Department of Natural Resources, and the McDonogh School all kindly provide access
or management of land used by the Baltimore Ecosystem Study for ecological,
hydrological, and meteorological field studies. Additional support and assistance has been provided by agencies, communities and individuals who are
specifically acknowledged in the datasets and publications summarizing work
they facilitated.
Special thanks are due to Beth Koda, David Tenenabaum, Stephen Kenworthy, Ian Yesilonis, Rose Williams and Greater Baltimore Area Master Gardeners
Larry and Vickie Klose, and Stuart Caplan who provided assistance in the
collection and processing of the data. Additional appreciation is extended to
Wanda MacLauchlan from the Maryland Cooperative Extension Service who
provided training to conduct the residential lawn care survey. The authors
would like to thank reviewers for their comments.
References
Band, L. E., Tague, C., Kenworthy, S. & Law, N. (2002) Coupling water, carbon and nitrogen cycling
with variable source area dynamics in forested and urbanizing catchments. Paper presented at the
American Geophysical Union 2002 Spring Meeting, Washington, DC.
Brady, N. C. (1990) The Nature and Property of Soils, 10th edn (New York: MacMillan Publishing
Company).
Center for Watershed Protection (2000) Nutrient movement from the lawn to the stream, Watershed
Protection Techniques, 2, pp. 239–246.
Erickson, J. E., Cisar, J. L. I., Volin, J. C. & Snyder, G. H. (2001) Comparing nitrogen runoff and
leaching between newly established St. Augustine grass turf and an alternative residential
landscape, Crop Science, 41, pp. 1889–1895.
Flipse, W. J. Jr., Katz, B. G., Lindner, J. B. & Markel, R. (1984) Sources of nitrate in groundwater in
a sewered housing development, Central Long Island, New York, Groundwater, 22, pp. 418–426.
Garn, H. S. (2002) Effects of lawn fertilizer on nutrient concentration in run-off from lakeshore lawns,
Lauderdale Lakes, Wisconsin. Water-Resources Investigations Report 02-4130. Middleton, Wisconsin: United States Geological Survey.
Gold, A. J., DeRagon, W. R., Sullivan, W. M. & Lemunyon, J. L. (1990) Nitrate-nitrogen losses to
groundwater from rural and suburban land uses, Journal of Soil and Water Conservation, pp. 305–
310.
Groffman, P.M., Pouyat, R. V. & Williams, C. O. (2002) Soil nitrogen cycling in urban forests and
grasslands. Paper presented at the American Geophysical Union 2002 Spring Meeting, Washington, DC.
Groffman, P. M., Law, N. L., Belt, K. T., Band, L. E. & Fisher, G. T. (2003) Nitrogen fluxes and
retention in urban watershed ecosystems, Ecosystems, 7, pp. 393–403.
Nitrogen Input in Watersheds from Residential Lawn Care
755
Grove, J. M., Cadenasso, M. L., Burch, W. R., Jr., Pickett, S. T. A., Schwarz, K., Wilson, M. A. &
Boone, C. (2003) In Review. An ecology of prestige: Association of Group Identity and Social
Status with Ecological Structure in the Baltimore Metropolitan Region, Maryland. Society &
Natural Resources.
Hamilton, G. W. & Waddington, D. V. (1999) Infiltration rates on residential lawns in central
Pennsylvania, Journal of Soil and Water Conservation, 54, pp. 564–568.
Kelling, K. A. & Peterson, A. E. (1975) Urban lawn infiltration rates and fertilizer run-off losses under
simulated rainfall, Soil Society of America Proceedings, 39, pp. 348–352.
King, K., Harmel, R. D., Torbert, H. A. & Balogh, J. C. (2001) Impact of a turfgrass system on nutrient
loadings to surface water, Journal of the American Water Resources Association, 37, pp. 629–640.
Law, N. L. (2003) Analysis of water quality trends in urban-suburban watersheds, Unpublished
Dissertation, University of North Carolina at Chapel Hill.
Legg, A., Bannerman, R. T. & Panuska, J. (1996) Variation in the relation of rainfall to run-off from
residential lawns in Madison, Wisconsin, Water-Resources Investigation Report 96-4194, Madison,
Wisconsin: US Geological Survey.
Line, D. E., White, N. M., Osmond, D. L., Jennings, G. D & Mojonnier, C. B. (2002) Pollutant export
from various land uses in the Upper Neuse Basin, Water Environment Research, 74, pp. 100–108.
Liu, H., Hull, R. J. & Duff, D. T. (1997) Comparing cultivars of three cool-season turfgrasses for soil
water NO ⫺ 3 concentration and leaching potential, Crop Science, 37, pp. 526–534.
Maryland Department of Planning (2001) Maryland’s Changing Land: Past, Present and Future (Baltimore, MD: Maryland Department of Planning).
Maryland Turfgrass Council (1996) 1996 Economic Survey of Turfgrass (MD, Maryland Turfgrass
Council).
Miltner, E. D., Branham, B. E., Paul, E. A. & Rieke, P. E. (1996) Leaching and mass balance of
15
N-labeled urea applied to a Kentucky bluegrass turf, Crop Science, 36, pp. 1427–1433.
Morton, T. G., Gold, A. J. & Sullivan, W. M. (1988) Influence of overwatering and fertilization on
nitrogen losses from home lawns, Journal of Environmental Quality, 17, pp. 124–130.
North Carolina State University (NCSU) Water Quality Group (2000) Nutrient Management Evaluation
in the Neuse River Basin 319(h) Project (Raleigh, NC: Department of Biological and Agricultural
Engineering).
Osmond, D. L. & Platt, J. (2000) Characterization of suburban nitrogen fertilizer and water use on
residential turf in Cary, North Carolina, Horticultural Technology, 10, April–June.
Petrovic, A. M. (1990) The fate of nitrogenous fertilizers applied to turfgrass, Journal of Environmental
Quality, 19, pp. 1–14.
Pouyat, R. V., Groffman, P. M., Yesilonis, I. & Hernandez, L. (2002) Soil carbon pools and fluxes in
urban ecosystems, Environmental Pollution, 116, pp. 107–118.
Pouyat, R.V., Yesilonis, I. & Russell-Anelli, J. (2004) Soil chemical and physical properties in an urban
landscape, Journal of Environmental Quality, (submitted).
Robbins, P., Polderman, A. & Birkenholtz, T. (2001) Lawns and toxins, Cities, 18, pp. 369–380.
Starr, J. L. & DeRoo, H. C. (1981) The fate of nitrogen fertilizer applied to turfgrass, Crop Science, 21,
pp. 531–536.
Swann, C. (2001) The influence of septic systems at the watershed level, Watershed Protection
Techniques, 3, pp. 821–834.
Teyssot, G. (1999) The American lawn: surface of everyday life, in: G. Teyssot (Ed.) The American
Lawn (New York: Princeton Architectural Press).
USDA (1976) Soil Survey of Baltimore County, Maryland (Washington, DC: United States Department
of Agriculture).
US Geological Survey (1999) The quality of our nation’s waters—nutrients and pesticides. US
Geological Survey Circular 1225, Reston, Virginia.
Vitousek, P. M., Aber, J. D., Howarth, R.W., Likens, G. E., Matson, P. A., Schindler, D. W.,
Schlesinger, W. H. & Tilman, D. G. (1997) Human alteration of the global nitrogen cycle: sources
and consequences, Ecological Applications, 7, pp. 737–750.
Vogel, M. (1999) Environmental Kuznet Curves: A Study on the Economic Theory and Political Economy
of Environmental Quality Improvements in the Course of Economic Growth (New York: Springer-Verlag).
Waschbusch, R. J., Selbig, W. R. & Bannerman, R. T. (1999) Sources of phosphorus in stormwater and
street dirt from two urban residential basins in Madison, WI, 1994–95. US Geological Survey
Water-Resources Investigations Report 99-4021, Madison, WI.
Weller, D. E., Jordan, T. E., Correll, D. L. & Liu, Z. J. (2003) Effects of land use change on nutrient
discharges from the Patuxent River watershed, Estuaries, 2A, pp. 244–266.